Google Launches Gemini 3.7 Flash With Coding Gains and Half Price Intro Rates

Summary

Google DeepMind introduced Gemini 3.7 Flash on August 13, 2026, three weeks after 3.6 Flash. Tulsee Doshi, Senior Director of Product Management, attributes it to developer feedback and algorithmic innovations intended for future models. Google calls it its most intelligent workhorse for coding and agents, citing higher first-pass accuracy and production-ready code: FrontierCode 1.1 Main reached 43.6% versus 34.4%, and DeepSWE v1.1 reached 65.3% versus 49.0%. WebDev Arena rose to 1588 from 1538 Elo, with more complete apps in fewer prompts and stronger reference-based UI adherence. GDP.pdf improved to 34.0% from 22.0%, while AutomationBench climbed to 30.4% from 17.0%. Google says 3.7 better handles roadblocks, intent, instructions, planning and tool calls, reducing oversight and retries. Demos paired it with Nano Banana for a real-time 3D game, Gemini Omni sub-agents for parallax pages, a 3 agent robotics training loop and PDF-to-interactive data stories.

Introductory pricing through December 31, 2026, is $0.75 per 1M input tokens and $3.75 per 1M output tokens, half 3.6 Flash’s original rate. On January 1, 2027, prices double to $1.50 and $7.50. Starting today, Gemini Spark, the 24/7 personal agent launched at I/O, uses 3.7 Flash for Google AI Pro and Ultra subscribers in more than 160 countries, improving Google Workspace tool use for file consolidation, email drafting and status updates. Developers can use Google Antigravity or the Gemini API through Google AI Studio and Android Studio. Enterprises get the Gemini Enterprise Agent Platform and app. Updated Frontier Safety safeguards address Chemical, Biological, Radiological and Nuclear misuse and cyber offense. Google provides a model card and cites stronger early customer results without naming customers or metrics.